News Summary

Microsoft, along with OpenAI, is probing allegations that Chinese AI startup DeepSeek improperly used OpenAI’s APIs to train its own AI models. According to reports, DeepSeek may have exfiltrated large amounts of data from OpenAI, potentially breaching the terms of service which prohibit using OpenAI’s API outputs to develop competing models or extract data programmatically. The investigation centers on whether DeepSeek circumvented OpenAI’s rate limits through a method known as distillation, allowing them to access and utilize data at scale. This case highlights concerns over data security and the potential legal ramifications for violating API usage agreements.